Trajectory tracking control of wheeled mobile manipulator based on fuzzy neural network and extended Kalman filtering

For robot trajectory tracking control, it is necessary to model inverse dynamics system sufficiently well to allow high-performance control. However, for complex robots such as wheeled mobile manipulators (WMMs), it is often difficult to model the dynamics system owing to system uncertainties, nonlinearity, and coupling. In this paper, we propose an effective tracking control method based on fuzzy neural network (FNN) and extended Kalman filter (EKF) to achieve WMM followed reference trajectory efficiently. The FNN is trained to generate a feedforward torque. In order to increase the computational efficiency and precision of the training algorithm, the EKF is used to sequentially update both the output weights and centers of the FNN. The effectiveness of the proposed control algorithm is confirmed through system experiments.